The Regulator of G Protein Signaling RGS4 Selectively Enhances α2A-Adreoreceptor Stimulation of the GTPase Activity of Go1α and Gi2α*

Agonist-stimulated high affinity GTPase activity of fusion proteins between the α2A-adrenoreceptor and the α subunits of forms of the G proteins Gi1, Gi2, Gi3, and Go1, modified to render them insensitive to the action of pertussis toxin, was measured following transient expression in COS-7 cells. Addition of a recombinant regulator of G protein signaling protein, RGS4, did not significantly affect basal GTPase activity nor agonist stimulation of the fusion proteins containing Gαi1and Gαi3 but markedly enhanced agonist-stimulation of the proteins containing Gαi2 and Gαo1. The effect of RGS4 on the α2A-adrenoreceptor-Gαo1 fusion protein was concentration-dependent with EC50 of 30 ± 3 nm and the potency of the receptor agonist UK14304 was reduced 3-fold by 100 nm RGS4. Equivalent reconstitution with Asn88-Ser RGS4 failed to enhance agonist function on the α2A-adrenoreceptor-Gαo1 or α2A-adrenoreceptor-Gαi2 fusion proteins. Enzyme kinetic analysis of the GTPase activity of the α2A-adrenoreceptor-Gαo1 and α2A-adrenoreceptor-Gαi2 fusion proteins demonstrated that RGS4 both substantially increased GTPaseV max and significantly increasedK m of the fusion proteins for GTP. The increase inK m for GTP was dependent upon RGS4 amount and is consistent with previously proposed mechanisms of RGS function. Agonist-stimulated GTPase turnover number in the presence of 100 nm RGS4 was substantially higher for α2A-adrenoreceptor-Gαo1 than for α2A-adrenoreceptor-Gαi2. These studies demonstrate that although RGS4 has been described as a generic stimulator of the GTPase activity of Gi-family G proteins, selectivity of this interaction and quantitative variation in its function can be monitored in the presence of receptor activation of the G proteins.

Initiation of signal transduction cascades involving heterotrimeric G proteins requires that the binding of an agonist ligand to a G protein-coupled receptor (GPCR) 1 results in the stabilization of conformations of the GPCR which increase the rate of dissociation of bound GDP from the nucleotide binding pocket of the ␣ subunit of a cognate G protein. Binding of GTP is thus allowed. With GTP bound the G protein produces a series of activating conformational changes and can thence regulate, directly or otherwise, the activity of several enzymes which generate intracellular second messengers or the probability of opening of a range of ion channels (1). Effector regulation is terminated by the intrinsic GTPase activity of the G protein ␣ subunit. Although this cycle of events is clear (1), it is only in recent years that explanations for the measured discrepancy in GTPase activity rates of heterotrimeric G proteins, which appeared too slow to account for biochemically and electrophysiologically measured functional end points, have begun to become apparent. A number of effector enzymes have been shown to display GTPase activating protein (GAP) activity toward their partner G proteins (2)(3)(4). Moreover, a relatively recently identified family of regulators of G protein signaling (RGS) proteins (see Refs. 5-8, for reviews) play key roles in accelerating the GTP hydrolysis rates of at least the G i and G q families of G proteins. RGS GAP activity is thought to facilitate attenuation of functional output of G protein-coupled signaling systems (9 -13). Data from both biochemical assays (14,15) and the crystal structure of the core catalytic domain of RGS4 bound to a transition-state model of G␣ i1 (16) has indicated this to be the key state for these interactions.
Many initial studies on the interactions of G proteins and RGS family members simply used "single turnover" studies to confirm RGS function (15,17). Such studies involve the coaddition of a recombinant RGS and a G protein which has been preloaded with [␥-32 P]GTP under ionic conditions which limit nucleotide hydrolysis. Release of this constraint allows measurement of accelerated GTP hydrolysis in the presence of an active RGS. Although highly useful, only a limited amount of kinetic information can be derived from such studies, and steady-state GTPase activity and its regulation by the presence of GPCRs and appropriate agonist ligands can provide more detailed information (4,11,18). However, to date, such studies have been restricted to the reconstitution of GPCR (e.g. the M1 and M2 muscarinic receptors), G protein, and RGS into phospholipid vesicles and have not been performed in native membrane systems under conditions capable of monitoring GTPase turnover number.
We have recently generated a series of fusion proteins between the ␣ 2A -adrenoreceptor and the ␣ subunits of the pertussis toxin-sensitive G proteins by the simple expedient of fusing the N terminus of the open reading frame of the G protein cDNA in-frame with C-terminal end of the GPCR cDNA from which the stop codon was removed (19 -23). These fusion proteins have a number of distinct properties ideal for quantitative functional analysis (see Refs. 24 and 25 for reviews). First, they ensure a 1:1 stoichiometry of expression of GPCR and G protein. This ratio establishes the validity of saturation ligand binding studies using [ 3 H]antagonist as a direct measurement of the level of expression of the G protein as well as the GPCR, which is difficult to achieve with independently coexpressed GPCRs and G proteins. Second, the fusion protein strategy ensures equivalent proximity of the GPCR to each G protein linked to it. Most importantly, the fusion proteins function as agonist-activated GTPases. Addition of agonist to membranes expressing such a fusion protein results in stimulation of GTPase activity (20,21). Although the two elements of such fusion proteins inherently cannot fully physically separate upon agonist stimulation, a considerable range of studies have recently questioned the previous assumption that such separation does indeed occur (Ref. 26, see Ref. 27 for review). Furthermore, a role for RGS proteins in the stabilization of GPCR-G protein interactions has also been suggested (4).
Many studies with closely related, isolated, G proteins have suggested that the effects of RGS proteins are generally nonselective. However, very recent work has indicated that the N-terminal region of RGS4 allows GPCR-selective regulation of signaling complexes which involve G q (28,29). In the current study we now demonstrate selective regulation by RGS4 of G i family G proteins coupled to the same GPCR and that the effect of RGS4 includes both concentration dependent increase in rate of agonist-stimulated hydrolysis of GTP and an increase in K m for this nucleotide.  (31). The amplicons generated using primers spanning the restriction sites DraI and EcoRI (respectively, in G␣ i1 and in pcDNA3) were subcloned into the ␣ 2A -adrenoreceptor-G␣ i1 fusion construct (20) restricted with the same enzymes. An identical strategy was applied to construct the pertussis toxin-resistant fusion protein ␣ 2A -adrenoreceptor-Val 351 -G␣ o1. The mutagenic reverse primer was designed to include the XhoI site to facilitate subcloning in pcDNA3, while the forward primer was designed to anneal to the sequence spanning the ClaI site in G␣ o1 . Equivalent strategies were used to generate ␣ 2A -adrenoreceptor-Ile 352 -G␣ i2 and ␣ 2A -adrenoreceptor-Ile 351 -G␣ i3 .

Materials-All
Cell Culture and Transfection-COS-7 cells were maintained in Dulbecco's modified Eagle's medium containing 10% (v/v) newborn calf serum, 2 mM L-glutamine. Cells were seeded in 60-mm culture dishes and grown to 60 -80% confluency (18 -24 h) prior to transfection with pcDNA3 containing the relevant cDNA species using LipofectAMINE reagent (Life Technologies, Inc.) (20). For transfection, 2.5-2.8 g of DNA was mixed with 10 l of LipofectAMINE in 0.2 ml of Opti-MEM (Life Technologies, Inc.) and incubated at room temperature for 30 min prior to the addition of 1.8 ml of Opti-MEM. COS-7 cells were exposed to the DNA/LipofectAMINE mixture for 5 h. 2 ml of 20% (v/v) newborn calf serum in Dulbecco's modified Eagle's medium was then added to the cells. Cells were harvested 48 h after transfection. In all the experiments herein cells were treated for the final 24 h prior to cell harvest with pertussis toxin (25 ng/ml).
Preparation of Membranes-Plasma membrane-containing P2 particulate fractions were prepared from cell pastes that had been stored at Ϫ80°C following harvest as described previously (32). ligand was separated from free by vacuum filtration through GF/C filters. The filters were washed with 3 ϫ 5 ml of assay buffer, and bound ligand was estimated by liquid scintillation spectrometry.
High affinity GTPase Assays-Were performed as described in Refs. 19 -21. Nonspecific GTPase was assessed by parallel assays containing 100 M GTP. All experiments were performed at least three times on membranes prepared from individual cell transfections.
Production of Recombinant RGS4 -Hexahistidine-tagged wild type and Asn 88 -Ser RGS4 were expressed in Escherichia coli and purified as described previously (33).

RESULTS
A fusion protein was constructed between the porcine ␣ 2Aadrenoreceptor from which the stop codon had been removed and a variant of G␣ o1 in which Cys 351 , the site for pertussis toxin-catalyzed ADP-ribosylation, was converted to Val. This resulted in production of a single open reading frame in which the N terminus of the G protein was attached in-frame with the C terminus of the GPCR. This construct was expressed transiently in COS-7 cells and following extensive treatment with pertussis toxin (25 ng/ml, 24 h) membranes were prepared. Addition of a maximally effective concentration of the ␣ 2 -adrenoreceptor agonist UK14304 (100 M) caused a substantial increase in high affinity GTPase activity measured using 0.5 M GTP as substrate. Parallel specific ligand binding studies were performed in these membranes with the ␣ 2 -adrenoreceptor antagonist [ 3 H]RS-79948-197 (19 -23, 34). As the fusion protein strategy defines a 1:1 ratio of expression of the ␣ 2adrenoreceptor and G protein, the stimulation of the GTPase activity by UK14304 at this concentration of GTP was calculated to be 2.8 Ϯ 0.4 mol of GTP hydrolyzed/mol of fusion protein Ϫ1 min Ϫ1 (mean Ϯ S.E., n ϭ 6). When equivalent studies were performed using a fusion protein between the ␣ 2A -adrenoreceptor and Val 351 -G␣ i1 , UK14304 (100 M) stimulation of the high affinity GTPase of this fusion protein was substantially greater at 8.5 Ϯ 0.3 mol of GTP hydrolyzed/mol of fusion protein Ϫ1 min Ϫ1 (mean Ϯ S.E., n ϭ 3). (Fig. 1). Addition of purified, recombinant RGS4 (33) (30 nM) to membranes expressing the ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 fusion protein did not significantly alter basal high affinity GTPase activity but produced a strong enhancement (p Ͻ 0.005) of UK14304 stimulated activity ( Fig. 1), increasing the effect of agonist to 10.0 Ϯ 0.2 mol of GTP hydrolyzed/mol of fusion protein Ϫ1 min Ϫ1 . In contrast to this effect of RGS4, addition of the recombinant protein to membranes expressing the ␣ 2A -adrenoreceptor-Val 351 -G␣ i1 fusion protein produced neither an alteration in basal GTPase activity nor a significant enhancement (p ϭ 0.42) of the stimulation of GTPase activity produced by UK14304 (9.1 Ϯ 0.3 mol of GTP hydrolyzed/mol of fusion protein Ϫ1 min Ϫ1 ) (Fig. 1).
Addition of varying amounts of recombinant RGS4 to membranes expressing the ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 fusion protein resulted in a concentration-dependent enhancement of the effect of UK14304 with an EC 50 for RGS4 of 30 Ϯ 3 nM (Fig. 2). This effect of RGS4 was likely to be dependent on specific, high affinity G protein binding since addition of equivalent amounts of an Asn 88 -Ser RGS4 mutant was unable to enhance the stimulation of GTPase produced by UK14304 (Fig.  2). This mutation produces an RGS4 protein which is unable to bind effectively to any form of G␣ i1 (33,35). Although RGS4 enhanced the effect of UK14304 on maximal stimulation of the GTPase activity of membranes expressing the ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 fusion protein, it also increased the EC 50 for the agonist by some 3-fold. In the absence of RGS4 this was 31 Ϯ 5 nM while in the presence of added RGS4 it was 113 Ϯ 23 nM (Fig. 3).
The ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 fusion protein can be effectively considered as a agonist-stimulated, single enzyme species. Basal and regulated GTPase activity in membranes expressing this fusion protein were measured at a wide range of concentrations of GTP to explore the kinetic basis of RGS4 action (Fig. 4). Analysis of this data by extrapolation to V max demonstrated that UK14304 increased the basal GTPase rate, as previously demonstrated for a number of other GPCR-G protein fusion proteins (20,36,37), and reduced K m for GTP (basal ϭ 0.36 Ϯ 0.04 M, UK14304 ϭ 0.21 Ϯ 0.03 M, mean Ϯ S.E. n ϭ 5, p ϭ 0.019). Addition of RGS4 (100 nM) along with UK14304 substantially further increased the V max of the system. Now, however, a marked increase (from 0.21 Ϯ 0.03 M in the absence of RGS4 to 1.26 Ϯ 0.16 M in the presence of 100 M RGS4, mean Ϯ S.E., n ϭ 5, p ϭ Ͻ 0.001) in the estimated K m for GTP was also observed (Fig. 4). With knowledge of the levels of expression of the ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 fusion protein from [ 3 H]RS-79948-197 binding studies, UK14304-stimulated GTPase turnover number at V max was calculated to increase from 6.3 Ϯ 1.4 min Ϫ1 to 81.7 Ϯ 16.0 min Ϫ1 with addition of 100 nM RGS4 (mean Ϯ S.E., n ϭ 5, p ϭ 0.002) ( Table I).
When varying amounts (1-100 nM) of RGS4 was added to membranes expressing the ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 fusion protein and kinetic analysis of UK14304 (100 M) stimulated GTPase activity was monitored, V max was progressively increased as was the measured K m for GTP (Fig. 5).

DISCUSSION
The expression of more than 20 members of the mammalian family of RGS proteins (5,6) suggests that they are likely to display marked selectivity of expression patterns and/or function. Striking variation in their distribution patterns in the central nervous system has been observed (38). By contrast, early studies of purified G proteins and RGS proteins indicated little selectivity in the capacity of individual RGS proteins to regulate the GTPase activity of individual members of the G i and G q family G proteins. However, emerging data suggest inherent selectivity of RGS proteins toward G proteins may indeed exist in their native environment. For example, a recent study examining the ability of a range of GPCRs to regulate Ca 2ϩ signaling via G q family G proteins in pancreatic acinar cells showed both marked variation in potency of individual RGS proteins to regulate function via a single GPCR and of the same RGS to regulate the signaling output from multiple, related, GPCRs (29). Although the basis of the interactions between the highly conserved core domain of RGS proteins with G␣ i1 has been elucidated at atomic level (16), recent studies have indicated that the N-terminal region of RGS4 might confer selectivity toward particular GPCR-G protein tandems (28). Such studies thus provide a conceptual basis to further explore selectivity in function of RGS proteins and suggest that direct interactions between RGS proteins and GPCRs occur.
Previously, selectivity in regulation of individual G i family G proteins by the most widely studied RGS, RGS4, has not been observed. Herein we demonstrate the selective enhancement of the GTPase activity of G o1 and G i2 by RGS4 when these G proteins are activated by the ␣ 2A -adrenergic receptor. Furthermore, for the first time, the experiments were performed in membrane preparations rather than in reconstituted lipid vesicles. These studies have utilized a series of fusion constructs between the ␣ 2A -adrenergic receptor and the ␣ subunit of each of G i1 , G i2 , G i3 , and G o1 in which the N terminus of the G protein was linked directly and in-frame with the C terminus of the GPCR from which the stop codon had been removed. We (19, 36. 37) and others (39 -42) have previously made considerable use of this strategy as it allows detailed enzyme kinetics to be performed on the G protein under conditions in which the ratio of expression of the GPCR and each individual G protein is kept constant (see Refs. 24 and 25, for reviews). Although the two elements of such fusion proteins inherently cannot fully physically separate upon agonist stimulation, a considerable range of studies have recently questioned the previous inherent implication that this should indeed occur (Ref. 26 All cell lines which are widely used for either transient or stable expression of GPCRs express a range of endogenous pertussis toxin-sensitive G proteins. Therefore, to ensure that agonist stimulation of high affinity GTPase activity following expression of the fusion proteins measured guanine nucleotide exchange and hydrolysis by the G protein linked to the GPCR we have used mutants of the G i family G proteins in which the Cys residue which acts as the acceptor for pertussis toxincatalyzed ADP-ribosylation was converted to either Val or Ile (31,34). Pertussis toxin-catalyzed ADP-ribosylation prevents effective interaction between modified G i family G proteins and GPCRs. Therefore, following expression of fusion proteins containing a mutation at this position, the cells were treated with pertussis toxin prior to membrane preparation and assay.  Table I. Stimulation of GTPase activity by agonist now represents only activation of the GPCR-linked G protein as G proteins of the G s , G q , and G 12 families, which are not sensitive to pertussis toxin, produce too limited a signal to be detected with this assay design, even if the receptor can interact productively with them. Initial studies expressed the ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 fusion protein in pertussis toxin-treated COS-7 cells. Parallel measures of the capacity of the agonist UK14304 to stimulate high affinity GTPase activity (when using 0.5 M GTP as substrate) and the levels of expression of the fusion protein indicated that a maximally effective concentration of UK14304 caused stimulation of hydrolysis of some 2-3 mol of GTP⅐mol of fusion protein Ϫ1 min Ϫ1 . Addition of recombinant RGS4 to these membranes increased the effect of UK14304 some 3-fold. Such results, although quantitatively more detailed than produced from previous studies, were essentially predictable from prior studies of RGS function. In addition, from isolated agonistinduced activation of fusion protein GTPase activity, we determined that the potency of UK14304 to stimulate the ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 fusion protein was slightly decreased by the presence of RGS4. Furthermore, the lack of ability of the Asn 88 -Ser form of RGS4 (33) at concentrations up to 1 M to enhance the effect of UK14304 demonstrated the specific requirement for a high affinity interaction between RGS4 and the ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 fusion protein.
When equivalent studies were performed using an ␣ 2A -adrenoreceptor-Val 351 -G␣ i1 fusion protein, however, very different results were obtained. First, UK14304 stimulation of the GTPase activity of ␣ 2A -adrenoreceptor-Val 351 -G␣ i1 at 0.5 M GTP, in the absence of added RGS4, was substantially greater than of ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 . More importantly, however, addition of recombinant RGS4 now failed to increase significantly the stimulation of GTPase activity produced by  Table I. UK14304. These data provide the first evidence that the capacity of a RGS protein to enhance the GTPase activity of a G i -family G protein might be dependent upon the GPCR which produces the primary stimulus of GDP release and subsequent guanine nucleotide exchange and also suggest that a direct contact between G i -coupled GPCRs and RGS proteins might occur. Such a mechanism is not without precedent: an alternatively spliced, larger form of the RGS protein RGS12 contains a PDZ motif which has been shown to bind certain GPCRs (43).
Analysis of UK14304 stimulation of the GTPase activity of ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 and ␣ 2A -adrenoreceptor-Val 351 -G␣ i1 at varying concentrations of GTP indicated K m for the substrate to be similar (Table I). The observed higher GTPase activity of ␣ 2A -adrenoreceptor-Val 351 -G␣ i1 /mol, when measured at 0.5 M GTP compared with the ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 therefore cannot simply reflect a substantially lower K m for substrate of this G protein. GDP release is well accepted to be the rate-limiting element of the guanine nucleotide exchange and hydrolysis which is promoted by receptor activation. It is thus important to note that G␣ o1 has a 5 times higher rate of basal GDP release than does G␣ i1 (44). This further indicates that the greater agonist stimulated GTP turnover by ␣ 2A -adrenoreceptor-Val 351 -G␣ i1 must reflect more effective activation of this G protein by the receptor.
Equivalent studies also demonstrated that the effect of RGS4 on the ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 fusion protein was to substantially increase V max of the GTPase activity. Most impressively, however, full kinetic analysis demonstrated that addition of RGS4 to membranes expressing the ␣ 2A -adrenoreceptor-Val 351 -G␣ o1 fusion protein resulted in a marked increase in the K m for agonist stimulation of GTPase activity which was dependent upon RGS4 levels. As such, assays performed at a single concentration of GTP substantially underestimate the capacity of RGS4 to enhance agonist-mediated GTP hydrolysis. Indeed, rather than the estimates of a 3-fold stimulation produced by different preparations of recombinant RGS4 when the assays were performed at 0.5 M GTP, extrapolation of the kinetic parameters to V max indicated the true stimulation to be some 15-fold.
If the role of RGS4 is to stabilize the transition state for GTP hydrolysis without altering GDP release or GTP loading, as has previously been surmised (16) these are exactly the kinetic characteristics expected.
An RGS4-induced increase in k 2 without alteration in the other kinetic constants will result in an increase in K m as, It was also the case that RGS4 increased both agonist-stimulated V max and the K m for GTP of the ␣ 2A -adrenoreceptor-G␣ i2 fusion protein (Table I). However, the effects of RGS4 were such that with 100 nM RGS4, a maximally effective concentration of agonist and measurement of GTPase activity at V max , the enzyme turnover number of this fusion protein was only 30% of that produced by the ␣ 2A -adrenoreceptor-G␣ o1 fusion protein under equivalent conditions (Table I). These result reinforce the quantitative differences in the effects of RGS4 on receptor regulation of closely related G proteins and also indicate why analysis at V max is vital to produce detailed insights. It was also noticeable that the addition of recombinant RGS4 did not significantly increase basal high affinity GTPase activity. This observation suggests "basal," RGS4 resistant, GTPase in such assays, generally considered to represent a composite signal which, at least in large part, represent the relatively high guanine nucleotide exchange of endogenous G i -family G proteins might actually be derived largely from non-G protein GTPases, such as tubulin, which contaminate the membrane preparations used for these studies. Alternatively, if high affinity interactions of RGS4 in native systems involve interactions with both GPCR and G protein (28) then pertussis toxinmediated uncoupling of endogenously expressed G proteins and GPCRs may encourage selective interactions of the added RGS4 with the expressed fusion protein.
If the differences in sensitivity of the GPCR-G protein fusion proteins to RGS4 are not due to a receptor interaction, an alternative explanation for its selectivity might lie in primary structure differences of the three switch regions of closely related G␣ i proteins, particularly in switch 3. These regions undergo conformational changes upon GDP-GTP exchange and are thought to be the principal contact points with RGS proteins (16). Interestingly, two-hybrid approaches using the RGS family member GAIP have shown strong interactions of this protein with G i1 , G i3 , and G o1 but only weak interactions with G i2 (45). This difference in affinity of interaction between G i1 and G i2 has been suggested to be due to the presence or absence of a single Asp residue in the Switch 3 region in G i1 . In G i2 the equivalent residue is an Ala (46). This may not represent the entire answer as G o1 , which interacts well with GAIP, has a Gly rather than Asp at the equivalent position. It should be noted, however, that G i1 and G i3 , which were not effectively regulated by RGS4 in this study, are the most closely related in sequence of the G i family G proteins. It will be of interest to ascertain if a similar pattern of selectivity for RGS4 is observed for activation of these G proteins by different GPCRs or if different members of the RGS family will display such selectivity.
Each of the GPCR, the G protein and the RGS4 are potential targets for post-translational S-linked palmitoylation (47,48). In the case of the ␣ 2A -adrenoreceptor, acylation occurs at Cys 442 (49). However, mutation of this residue to Ala does not intefere with the effectiveness of coupling of this receptor either in co-expression studies (49) or following construction of this mutant into an ␣ 2A -adrenoreceptor-G i1 ␣ fusion protein (19). G␣ i1 is palmitoylated at Cys 3 (50) as are the other G i family ␣ subunits. Although there is no formal proof that this amino acid becomes palmitoylated in the fusion proteins used herein the equivalent Cys residue does become palmitoylated, and can be regulated in an agonist-dependent manner, in a ␤ 2 -adrenoreceptor-G␣ s fusion protein (51). This does not appear directly relevant for the the present studies, however, as it has previously been shown that an ␣ 2A -adrenoreceptor-G␣ i1 fusion protein in which this Cys is substituted by Ala allows as effective agonist stimulation of the G protein GTPase activity as in a wild type fusion protein (19). There has been considerable recent interest in observations that a range of RGS family members including RGS4, RGS10 (52), and RGS16 (53) can be palmitoylated. A palmitoylation-defective mutant of RGS16 is able to act as an RGS for G i but following expression in HEK293 cells was impaired in its capacity to attenuate both G iand G q -mediated signal cascades (53). In the case of RGS4, mutants designed to limit palmitoylation had different effects on the GAP activity of the protein against G i dependent upon whether the assay was based on "single-turnover" measurements of GTP hydrolysis or in steady state, ligand-regulated, GTPase assays which are closer to the assay system we have employed. In this system palmitoylation promoted GAP activity (52). The bacterially produced recombinant protein we have used herein will not be palmitoylated and we may thus have underestimated the absolute GAP capacity of RGS4 in our assays. As such, this will be an important issue to be addressed in future studies. However, there is no current evidence to suggest that the acylation status of an RGS protein will differentially modulate its capacity for regulation of a series of closely related G proteins which is the key observation in the current studies.
It has been well established in a range of systems that the ␣ 2A -adrenoreceptor can concomitantly interact with and activate each of the G i -family G proteins (54,55). As such activation can produce regulation of effector systems as diverse as adenylyl cyclase (54), ERK MAP kinases (56), voltage-operated Ca 2ϩ channels (57), and K ϩ channels (57), then the ability of RGS4 to selectivity control the duration of action of these ␣ 2A -adrenoreceptor-activated G proteins is likely to allow distinct regulation of different end points.